Speaker
Description
The development of longitudinal diagnostics for short-pulse electron accelerators is challenging but necessary to provide high-brightness electron bunches. This is equally true for novel plasma accelerators as for free electron lasers. The gold standard for such measurements is a transverse deflecting cavity (TDC), but these are typically invasive, are costly to produce and operate, and have resolution limited to around 10 fs. To this end, a THz-based reflective imaging system has been designed and installed at the MAX IV Short Pulse Facility (SPF) for imaging Coherent Transition Radiation (CTR). This contribution presents the initial results of applying transfer learning to deep learning models, such as convolutional neural networks, for evaluating the reconstruction of bunch profiles from experimentally acquired CTR images, building on previous successes in profile prediction using simulated data. The reconstruction of the longitudinal profile can be achieved using single CTR images and benchmarked against a TDC. Practical resolution limits and the next steps in the development of this monitor are also discussed.
Funding Agency
EU’s Horizon Europe research and innovation program no. 101073480 and UKRI guarantee funds, Cockcroft Institute core grant no. STFC ST/V001612/1, STFC AWAKE-UK phase II grant no. ST/X005208/1.
| I have read and accept the Conference Policies | Yes |
|---|
